Optical coherence tomography of fovea before and after laser treatment in retinopathy of prematurity.

PURPOSE: To study the fovea in preterm babies with Type I retinopathy of prematurity (ROP) before and after laser treatment using optical coherence tomography (OCT).
MATERIALS AND METHODS: This was a prospective observational case-control study including preterm neonates undergoing screening for ROP from May 2009 to July 2011. Group 1 included 30 eyes of 15 neonates with Type I ROP. A 532-nm laser was used for treatment in all cases for Group 1. Group 2 included 14 eyes of 7 preterm neonates without ROP that served as controls. OCT was performed under sedation in the lateral position before and after laser treatment. P <0.05 was considered statistically significant.
RESULTS: The mean initial central macular thickness (CMT) was comparable in both groups (P = 0.832) and statistically significantly correlated with gestational age (P = 0.015). No adverse effects on the anterior segment or posterior segment were observed. There was no significant difference in CMT before and after laser treatment in Group 1 (P = 0.007). There was one case of cystoid macular edema after laser treatment.
CONCLUSION: The macula in preterm babies with Type 1 ROP was comparable to those without ROP. Gestational age was the only predictor of CMT.

INTRODUCTION

Retinopathy of prematurity (ROP), previously known as retrolental fibroplasia (RLF)1 is a disease of the eye that affects primarily premature and low birth weight babies. The incidence of ROP in the developing world is increasing and reported to vary between 38% and 51.9%.23

Indirect ophthalmoscopy has been used successfully to diagnose various stages of ROP and to treat ROP.4 However, fine macular details which are not visible with indirect ophthalmoscopy can be visualized by optical coherence tomography (OCT). OCT is a noninvasive, high-resolution imaging modality used to evaluate macular anatomy. It has been shown to be comparable to in vivo histology. However, data are rare on OCT in ROP. There is increasing interest in OCT for ROP where detailed contact lens examination is not possible. It is seen that long-term visual acuity may be highly variable in patients with a history of retinopathy of prematurity (ROP),5 even when macular anatomy appears ophthalmoscopically normal. With increasing survival of premature infants, a better understanding of the anatomy of these eyes may give insight into the reasons for the variability in their visual acuity on long-term follow-up despite apparently normal macular anatomy.

The present study aims at studying the OCT imaging features of the macula in preterm babies with Type 1 ROP before and after 532-nm laser therapy. The OCT findings are also compared with a control group of preterm babies without ROP.

MATERIALS AND METHODS

This was an institution-based, prospective, control trial. The study was approved by the institutional review board and ethics committee. All the babies with birth weight less than 2000 g and gestational age less than 35 weeks are routinely screened for ROP. The study included 44 eyes of 22 premature babies screened for ROP. Thirty eyes of 15 consecutive babies that developed Type 1 ROP (high risk prethreshold ROP as per ETROP4) constituted Group 1 of the study. Fourteen eyes of seven preterm neonates, who were randomly selected from amongst screened babies, served as controls and constituted Group 2. All babies in Group 2 had reached approximately 40 weeks of gestation, without developing any signs of ROP.

We excluded babies with: (1) Stage 5 ROP; (2) Vitreous hemorrhage; (3) Persistent tunica vasculosa lentis or nondilating pupil; (4) Babies in incubators or sick babies where OCT scan was not possible; (5) Babies with systemic problems where proper positioning for OCT was not possible or when midazolam was contraindicated.

We reviewed the neonatology records for gestational age, birth weight, post-conceptional age, neonatal problems like respiratory distress, hypoglycemia, seizures, dyselectrolytemia, neonatal jaundice, use of oxygen after birth, apnea, transfusion of blood and blood components, sepsis, intracranial hemorrhage, use of surfactant, oxygen saturation, metabolic acidosis, and respiratory acidosis.

The pupils were dilated with topical 2.5% phenylephrine and 0.5% cyclopentolate. A single drop of each was instilled twice at 20-minutes interval. The screening was performed with indirect ophthalmoscopy and a + 28 D lens. Any baby with ROP underwent a detailed examination using an eye speculum and wire vectis for globe rotation. The exact zone, stage, and extent of ROP and presence of plus disease were recorded.

OCT was performed using time domain OCT (TD-OCT) (Stratus OCT Model-3, software 4.0.7; Carl Zeiss, Jena, Germany) with the infant sedated with intravenous midazolam (0.05-0.1 mg/kg body weight) and triclofos oral syrup (10-15 mg/kg). OCT was performed in the lateral position by positioning the patient table adjacent to the OCT machine. An eye speculum was used to open the eye and a wire vectis was used as required for proper positioning of the globe. When visible, the foveal reflex was centered for imaging. As babies in this age group often do not have a clearly identifiable foveal reflex, this was not always possible. In these cases, the anatomic relationship of the optic nerve head and vascular arcades was used as a landmark for imaging and the fast macular scan was used. OCT scans were used to study the foveal thickness, foveal dip, microcystic changes, macular edema and subretinal/intraretinal fluid. All examinations and treatments were performed in the presence of a pediatrician (SJ, DC). In cases with a pitless fovea, OCT was used to scan through the whole macular region by raster lines to confirm the absence of the pit. The examination and OCT was repeated 1 week after laser treatment in Group 1. In Group 1, the babies were treated using 532-nm YAG Laser Indirect Ophthalmoscope (LIO). The laser power, the duration of burns and the number of burns were recorded. The mean power used was 150 ± 40 mW for 150 ± 20 msec with the mean of 1232 ± 570 burns. During treatment, near confluent burns were applied to the avascular retina. Repeat laser was performed for any skipped areas at the 1-week follow-up. ROP regressed in all but one baby who had aggressive posterior retinopathy of prematurity (APROP). He developed stage 4A ROP despite laser treatment and later underwent lens sparing vitrectomy. Although the OCT unit was not calibrated for neonatal eyes, the foveal thickness was compared between two groups and, within Group 1, before and after the laser procedure, to look for any obvious variation. In the babies showing macular edema, OCT scans were repeated monthly till near normal foveal anatomy was achieved.

Statistical analysis

Statistical analysis was performed with SPSS 13.0 package (IBM Corp., New York, NY, USA). For continuous data, descriptive statistics such as mean, mode, median, standard deviation and range were calculated. Data were checked for skewness. For skewed data, the Mann-Whitney test was applied. The Wilcoxon signed rank sum test was applied for repeated measure continuous data. For deriving the correlation in different parameters, Spearman's correlation coefficient analysis was used. A P value less 0.05 was statistically significant.

RESULTS

In the study period from May 2009 to July 2011, 162 inborn babies were screened for ROP. Of these babies, 17 developed Type 1 ROP (treatable ROP). Two babies were excluded from the study as OCT could not be performed on them as they were in incubators. Group 1 included 30 eyes of 15 premature babies with Type 1 ROP and Group 2 included 14 eyes of seven premature babies without ROP.

In Group 1 and 2, the mean birth weight was1285.8 ± 313.9 g and 1278.6 ± 237.8 g, respectively, the gestational age was 30.5 ± 2.2 weeks and 31.9 ± 2.5 weeks, respectively, and post-conception age (PCA) at first OCT examination was 35.1 ± 1.5 weeks and 37.3 ± 2.1 weeks, respectively [Table 1]. All eyes in Group 1 had plus disease. The extent of ROP was, zone I disease in four eyes and zone 2 in 26 eyes, stage 2 in 11 eyes and stage 3 in 19 eyes. The mean central macular thickness (CMT) as calculated manually from the scans was comparable in both the groups (201.6 ± 41.3 μm in Group 1 versus 202 + 44.5 μm in Group 2; P = 0.832). Foveal dip was present in 15 eyes (50%) and absent in 15 eyes (50%) before laser therapy in Group 1. In Group 2, foveal dip was present in 6 eyes (43%) and was absent in 8 eyes (57%) which was comparable to Group 1 (P = 0.783). Pre-laser OCT scan showed unilateral macular edema in two eyes of two babies. The babies were treated with laser therapy in both eyes and began topical steroids and cycloplegic drugs. There was regression of plus disease clinically and resolution of edema as seen on OCT, 1 week after laser treatment [Figure 1].

Table 1

Demographic profiles of children retinopathy of prematurity and a control groups

Figure 1

(a) Thirty-one weeks gestation age and 36 weeks of post-conception age baby with zone II, stage 3 ROP. Pre-laser OCT scan shows macular edema (b) One week post-laser scan of the eye in Figure 1 A showing shallow foveal dip

One week post-laser, OCT scans in Group 1 (PCA 33-38 weeks) showed no significant change from prelaser scans (mean CMT of 203.3 ± 33.8 μm). Foveal dip was present in 18 (60%) eyes. The disease regressed by 4 weeks in 28 eyes of 14 babies. However, repeat laser for skipped areas was required in four eyes of two babies. In two eyes of one baby with APROP, the disease progressed to stage 4A despite laser treatment.

Two eyes of one baby with stage 3, zone II ROP developed cystoid macular edema as seen on OCT 1 week after laser treatment [Figure 2]. Cystoid spaces resolved 6-8 weeks after laser treatment with the use of topical steroids and ketorolac. Two eyes of one baby with APROP showed preretinal hyper-reflective membrane like structure on OCT, 1 week after laser treatment [Figure 3]. This baby had progressive ROP despite laser treatment and subsequently developed stage 4A disease in both eyes and underwent lens sparing pars plana vitrectomy.

Figure 2

(a) Thirty weeks gestational age, 36 weeks post-conceptual age baby with zone II stage 3 ROP. One week post-laser OCT scan shows cystoid macular edema (b) Eight weeks post-laser OCT scan of the same baby showing resolution of cystoid spaces and shallow foveal dip

Figure 3

(a) Fundus picture of eye with aggressive posterior ROP showing stage 2 zone I ROP before laser (b) One week post-laser OCT scan of the eye in Figure 3 A after laser treatment showing premacular membrane

There was a statistically significant correlation between the pre-laser CMT in both groups to the gestational age (P = 0.015). CMT did not correlate with any other baseline variable including birth weight and extent of the disease (P > 0.05). Pre-laser CMT correlated statistically significantly with post-laser CMT (P = 0.007). There was no correlation between the extent of ROP, number of laser spots and post-laser foveal thickness (P > 0.05). Post-laser cystoid macular edema did not correlate to any of the baseline variables, extent, and location of ROP or number of laser spots (P > 0.05). No baby developed any sedation-related complications or required any supplemental oxygen or showed any change in oxygen saturation during or after laser treatment.

DISCUSSION

Fundus examination with indirect ophthalmoscopy is a widely for screening and planning treatment of ROP in preterm babies. OCT imaging has enhanced our understanding and has improved the treatment of adult macular diseases. However, positioning and cooperation requirements limit neonatal imaging with the routinely available tabletop OCT systems. Portable OCTs with hand held scanners are now widely used for neonatal imaging.6 However, the significant cost of the newer machines with hand held OCTs precludes availability in most institutes in developing countries. It is well documented that OCT monitoring of macula provides significant additional information which may have a bearing in the management and the outcome of neonates with ROP. Thus the simple procedure of using a table top TD-OCT for neonates by putting the baby in lateral position seems to be an economically viable option for evaluating the macula of ROP babies.

The major limitation of the use of TD-OCT is that it is not calibrated for use in neonates. The eyes of neonates have a smaller antero-posterior diameter and steeper corneas. The readings for TD-OCT in neonates have to be adjusted for the magnification factor. The CMT readings may not be accurate with TD-OCT. However, the present study shows that the scans seem to give us a fair idea about the macular status in neonates.

We found a statistically significant correlation between the pre-laser CMT in both groups to the gestational age (P = 0.015). Similar findings have been found in another study in which central macular thickness correlated with gestational age.7 In another study of older children with a history of threshold ROP, abnormal foveal contour and thicker foveas were seen in preterm babies with a history of threshold ROP.89 These babies had poorer visual outcome compared to children who had been term deliveries.8 In the present study, the foveal dip was absent in 50% of the eyes with ROP. However, this was comparable to the control group in which 57% of the eyes did not have a foveal dip. Some studies have shown a significant effect of ROP severity on OCT thickness.10 The influence of the severity of ROP on foveal thickness cannot be discussed from the data in our study due to low number of babies with severe disease (APROP).

Pre-laser macular edema was seen in two (6.7%) of the 30 eyes in the present study. In earlier studies subclinical CME including serous detachment has been as high as 58% in premature infants with ROP.111213 The discrepancy in the incidence between studies can be explained by the fact that the latter study by Maldonedo et al. included relatively younger babies that were undergoing screening for ROP. The present study included babies who had developed Type 1 ROP and were scheduled for laser treatment of the disease. OCT scans were performed just before laser and 1 week after laser treatment. Pre-laser macular edema as well as plus disease resolved after laser treatment. It is hypothesized that this could be associated with changes in VEGF levels in these eyes which are increased in vascularly active ROP and decrease after laser treatment.14

One week after laser treatment, clinically unapparent cystoid macular edema was diagnosed on OCT in one baby. The cystoid macular edema resolved after 8 weeks of topical steroids and the nonsteroidal anti-inflammatory drug ketorolac suggesting an inflammatory origin. Increased levels of inflammatory mediators such as interleukins (IL6) have been reported after laser treatment.15 Mulvihill et al. reported bilateral serous retinal detachment in a boy with ROP immediately following diode laser and resolution of the same at 3 weeks with pigmentary macular changes.13 The YAG 532-nm laser could also caused inflammatory macular edema after laser treatment. Of the 15 babies, we found post-laser CME in one eye which resolved on treatment resulting in a normal appearing macula with foveal dip on OCT. The point of interest is that 8 weeks is not the usual duration for which post-laser ophthalmic topical medication is instituted. Thus OCT has a role after laser treatment for all ROP babies so that the duration of post-laser treatment can be ascertained.

The eye with preretinal membrane on OCT had progressive disease despite laser treatment which required further surgery. A recent histopathological study shows that fibroblasts and myofibroblasts are the main cells of the epiretinal membrane (ERM) in ROP. The glial cells play an important role in the progression of ROP.16 Thus if ERM is present on OCT in ROP babies, we may be able to predict the possibility of surgical intervention in these babies.

OCT may be useful in the diagnosis and management of clinically unapparent macular pathologies in ROP. These may have long-term implications on the visual status of the baby. TD-OCT could be used to subjectively evaluate the fovea of preterm babies before and after laser treatment to detect clinically unapparent and subtle but significant changes. These babies may show pre-laser macular edema, absent foveal dip, pre-retinal membrane which could have a bearing on the future visual development and management of these babies. After laser treatment we need to monitor the fovea of these babies as they could develop macular edema requiring treatment for a relatively longer time.

The major limitation of the present study is the small sample size and duration of measurements with TD-OCT that can lead to a decentration artifact.